DiyLightAnimation
Hardware => Lynx EtherDongle => Topic started by: Dennis Cherry on December 27, 2012,
-
For the past 2 years a couple of my sequences did not display correctly, It took me this year to get answers to figure out what was causing the problem.
Using LOR S3 and either the mini director and a computer the results where the same, The second year even converted that half of the display to DMX, still had problems I did not get my questions answered until the display was removed in 2011.
This year went 100% DMX using the etherdongle using LOR S3, guess what? still was not working no matter what I tried to change.
About ready to give up and decided to look at the actual data being generated for output, used a program called SACNview. Looking at the data using S3 in the sequence sections that did not look right showed nothing unusual.
Decided to try another program called HLS and imported my problem sequences and what to my wondering eyes did I see but the sequences are performing correctly. Went back and looked again at the data in SACNview and it looked OK using HLS but the LOR Frames Per Second varied from 40 to 36.8 FPS but in HLS I got a steady 40FPS. Tried this on 2 different computers.
So my question is does the etherdongle handle a changing frame rate or is the problem with S3 not outputting all the data?
If you are interested in knowing what part of the sequence would not work correctly, it is doing random ON/OFFs in 25 or 50 ms intervals over 8 or 16 channels. The LOR S3 looked like it was missing 2-3 frames sometimes.
Even the rest of the display looks so much smoother, the CCR Circles really POP now. I have had spectators even comment on the difference.
BTW: One of the problem sequences is Richard Holdman's "Amazing Grace" the fast flash part never looked right, now you have close your eyes when it starts the fast flash.
-
Dennis I know there are some much smarter guys on here that will most likely be able to give you better answers but I do know the Etherdongle outputs at 50ms so I would presume that trying push anything faster through it might account for missed frames but like I said others would have to answer the exact answer. I know for my show I used LSP 2.5 to sequence and when I tried to use the LSP scheduler to play my show it played OK but didn't look right so I converted all my sequences to X-light sequences and used X-Lights and what a difference it made, it was like a totally different show. SACNview is very useful tool along with Wireshark if you using E1.31 to track packets.
-
Dennis I know there are some much smarter guys on here that will most likely be able to give you better answers but I do know the Etherdongle outputs at 50ms so I would presume that trying push anything faster through it might account for missed frames but like I said others would have to answer the exact answer. I know for my show I used LSP 2.5 to sequence and when I tried to use the LSP scheduler to play my show it played OK but didn't look right so I converted all my sequences to X-light sequences and used X-Lights and what a difference it made, it was like a totally different show. SACNview is very useful tool along with Wireshark if you using E1.31 to track packets.
Thanks for your reply.
This problem was reported to LOR this year and they knew the Etherdongle was being used. The question presented was "what is the maximum timing to be used with the Eetherdongle" and the answer came back what RJ gave us which is 23ms, that is about 44 FPS. So 25ms (40 FPS) should work fine and it dose.
Also tried XLights and it did not fix the problem either, it was slightly better than S3.
I think that if running the DMX protocol, the timing is critical, that is keeping everything within a constant timing frame, LOR does not do that and you cannot setup a 23 or 25 ms timing slots, only 20,30,40,50ms etc. variable is out also LOR in chases allows you to violate the timing marks when doing chases. That may of worked fine using the LOR protocol only but I feel that does not work for DMX. I know HLS looks at the LOR and Vixen events and makes a decision if the cell has an event or not.
Maybe someone else here that knows the DMX standards can add more information on this. I am no expert on this but have done a lot of reading trying to get my display to work.
I do feel if specification on products is not published then it is easy to break something and then blame the company for having a bad product.
My past lives was being a Manufacturing Product Engineer and worked closely with UL, CSA, BABT, etc agencies, but again I am no expert on this.
-
I have Amazing Grace and running S3 with and Etherdongle. I'm going to have to go and watch it to see if there is anything different. So far the only thing I have noticed to switching to DMX is I think my lights match the music better.
-
You have me curious now. What would I need to do the test my S3 sequence using HLS. I have never looked at it so I'm clueless.
I watched my last take of Amazing Grace when I was using LOR controllers under the LOR protocol and it looks good but I'm viewing a 2" wide video vs watching my live display.
-
I did do some testing with a sequence using a timing grid of .05 and smaller using on/off/on/off ..... and LOR with the ED worked fine at .05. Anything less and it wasn't keeping up and was missing events.
I tried to figure out how to convert a LOR sequence to HLS and setup with the ED but I had no luck. I'm happy with what I have but I would like to see if I'm missing something on playback.
Thanks for the feedback.
Yes I think going to 25 ms is where S3 fails.
If you want to know how to get HLS running please PM me, it is not hard.
-
That would make sense as I was looking at the specs for DMX512 and the minimum update time for 513 slots is 22ms So moving down to 25ms timing is pushing th limits (as least as I understand it).
-
Technically either LOR S3 or the Etherdongle (ED) in conjunction with S3 is not working as Dennis has stated. With the help of Dennis I converted my Amazing Grace sequence to run under HLS and the fast on/off/on/off he stated is AMAZING under (HLS + ED) compared to (S3 + ED).
I'm interested to figure out what the issue really is. I'm willing to purchase a different E1.31 adapter like the j1sys EGC-D2 if it would address the issue or even help figure it out.
My guests don't know the difference, but I do now. :(
-
My guess is LOR S3 does not have the I/o routines maximized to output the data fast enough as HLS does. It would be interesting to see if another E1.31 dongle made a difference, but I am guessing Joe has done his homework and maximized output routines in HLS.
-
Technically either LOR S3 or the Etherdongle (ED) in conjunction with S3 is not working as Dennis has stated. With the help of Dennis I converted my Amazing Grace sequence to run under HLS and the fast on/off/on/off he stated is AMAZING under (HLS + ED) compared to (S3 + ED).
I'm interested to figure out what the issue really is. I'm willing to purchase a different E1.31 adapter like the j1sys EGC-D2 if it would address the issue or even help figure it out.
My guests don't know the difference, but I do now. :(
I too would like to see how the other E1.31 devices compare in this software test.
-
Dennis you might want to contact Ed at j1sys and see if you could borrow one since he is just up the road from you.
-
I'm willing to purchase one if there is a chance it might work. I'm not interested in switching software packages but I liked the results I seen.
I'm just not sure which model is the best for me. Right now I only need 2 universes of DMX so the smaller might be OK and be more flexible.
http://www.j1sys.com/ecg-d2/
ECG-D2
ECG-D4
ECG-DR4
-
I tested LOR S3 with an Enttec Pro both internal set at 40 FPS and 0 or unlimited which is 850fps. This was just by watching the LED's on and LE but I could see that "it" (LOR, not the LE) was missing frames when the the timing grid was less than .04. This leads me to believe it's not a hardware issue.
My guess is any software that outputs DMX at less than 25ms could miss frames. The part I don't understand is I don't think xLights even outputs DMX at the 25ms rate.
I find it hard to believe that more people haven't noticed this. How can a program truly support DMX properly that does not spew data fast enough?
What don't I understand? There must be something as there are far smarter people out there than I when it comes to DMX. :)
-
I tested LOR S3 with an Enttec Pro both internal set at 40 FPS and 0 or unlimited which is 850fps. This was just by watching the LED's on and LE but I could see that "it" (LOR, not the LE) was missing frames when the the timing grid was less than .04. This leads me to believe it's not a hardware issue.
My guess is any software that outputs DMX at less than 25ms could miss frames. The part I don't understand is I don't think xLights even outputs DMX at the 25ms rate.
I find it hard to believe that more people haven't noticed this. How can a program truly support DMX properly that does not spew data fast enough?
What don't I understand? There must be something as there are far smarter people out there than I when it comes to DMX. :)
I have already discussed this with DEVMIKE (Programmer) at LOR last week, Devmike was kind of vague in his response back about how LOR handles output data in DMX mode, but said he would bring this up in a staff meeting. Do not know anymore than that.
-
I did some more testing with sACNViewer using LOR S3, xLights and HLS from a FPS standpoint and my results are below.
| 15.6 - 16 xLights |
| 42.8 - 43.6 LOR S3 |
| 40 - 40.40 HLS |
Based on the FPS data alone I would conclude that:
- HLS and LOR should have a very responsive and smooth output
- xLights would be jerky
What I actually think that is happening is:
xLights and HLS are outputting more true data and S3 is hit and miss if this makes sense. I don't think it's a matter of FPS or S3 should be making xLights and HLS look poor but this is not the case.
-
The other piece of the puzzle you might want to check with Wireshark is what are the drop packet rates between the three. In that if LOR S3 is sending out a ton of packets BUT many are getting dropped or are bad packets while HLS is sending fewer packets but they are all going to where there supposed to go with no corruption then Then that MAY be the reason HLS has a lower frame rate but better response from the controllers.
-
The other piece of the puzzle you might want to check with Wireshark is what are the drop packet rates between the three. In that if LOR S3 is sending out a ton of packets BUT many are getting dropped or are bad packets while HLS is sending fewer packets but they are all going to where there supposed to go with no corruption then Then that MAY be the reason HLS has a lower frame rate but better response from the controllers.
Good point, will try that.
Here is some SACNview files of the data you will see HLS data is doing better than S3 or XLights.
https://dl.dropbox.com/u/23140982/Xlights%20Amazing%20Grace%202012e.zip